shenmbsw / tensorflow-yolo3

implement yolo3 on DETRAC

tensorflow-yolo3

Setup environment

1. virtualenv --python=python3.5 py35
2. source py35/bin/activate
3. cd [dir of tf-yolo3]
4. pip install -r requirements.txt

---

Data Prepare

1. Download the DETRAC dataset from detrac_website
2. chage the config.py file to the location of your downloaded dataset.
3. run script in data to parse XML file.

cd datasets
python XML_parser.py
python image_migrate.py  

Training

1. If you want to use original pretrained weights for YOLOv3, download from darknet53 weights
2. rename it as darknet53.weights, and modify the darknet53_weights_path in the config.py

wget https://pjreddie.com/media/files/darknet53.conv.74`  

3. Modify the data augmentation parameters and train parameters
4. Run yolo_train.py

Evaluation

1. Modify the pre_train_yolo3 and model_dir in config.py
2. Run detect.py

python detect.py --image_file ./test.jpg

Notice

If you want to modify the Gpu index, please modify gpu_index in config.py

Credit

@article{yolov3,
	title={YOLOv3: An Incremental Improvement},
	author={Redmon, Joseph and Farhadi, Ali},
	journal = {arXiv},
	year={2018}
}

Reference

• keras-yolo3
• tf-yolo3